16
Marital Status and Mental Health
Kristi Williams, Adrianne Frech,
and Daniel L. Carlson
Williams, Frech, and Carlson examine the evidence for an effect of marital sta-
tus on mental health, with a particular focus on the factors that identify who
benefits from marriage, who suffers from marital dissolution, and under what
circumstances. They evaluate three possible explanations for observed associa-
tions of marital status with mental health:(1) the marital resource model,(2) the
marital crisis model, and (3) selection bias. They conclude that the best recent
evidence suggests that, on average, entering marriage improves mental health and
exiting marriage undermines mental health, at least in the short run. However,
their central argument is that these average associations obscure a great deal of
heterogeneity in the experience of marriage and in its consequences for mental
health. The authors consider a range of individual, demographic, and relationship
characteristics that likely moderate the effect of marriage and marital dissolu-
tion on mental health. These include gender, marital quality, age/life course,
race/ethnicity, values and beliefs, and prior mental health. What other factors
likely influence whether marriage and divorce are beneficial, neutral, or harmful
for mental health? How might the impact of marriage and divorce on mental
health change with the times, particularly as alternative family forms become
more prevalent?
Introduction
A general consensus exists among social scientists and the public at large that mar-
riage provides substantial benefits to mental health. For many years, this conclusion
was based on cross-sectional studies comparing the average mental health of the
married to that of the unmarried at a single point in time. This research clearly
showed that married individuals report lower average levels of depression, psy-
chological distress, and psychiatric disorder and higher levels of life satisfac-
tion and subjective well-being (see Umberson & Williams, 1999, and Waite &
Gallagher, 2000, for reviews) than the unmarried. The consistency and relatively
large magnitude of observed differences, as well as their persistence across time
and in numerous countries (Mastekaasa, 1994; Stack & Eshleman, 1998), led to
the conclusion that marriage improves mental health for most people.
Research findings about marital status differences in mental health strongly res-
onate with cultural views about the individual and societal importance of marriage.
Perhaps as a result, they are frequently heralded by the news media with headlines
This chapter was prepared with support from Grant R01 HD054866–01A1 from the National Institute
of Child Health and Human Development, National Institutes of Health.
306
Marital Status and Mental Health 307
like “Stressed Out? Say ‘I Do’”(2005),“Lonely? Feeling Low? Try a walk...
Down the Aisle”(2006), and “Feel Blue? Say ‘I Do’”(2006). Such generaliza-
tions–both in the popular press and, to some extent, in the scientific literature–
suggest that marriage is a panacea for a range of individual and societal ills. They
also have real consequences for the lives of individuals. For example, social sci-
ence research on average marital status differences in mental health is commonly
cited in support of government policies that encourage disadvantaged single moth-
ers to wed (Maher, 2005; Nock, 2005; Rector, Pardue, & Noyes, 2003). Further,
some argue that the universality of marriage’s mental health benefits is overstated,
which contributes to the stigmatization of unmarried adults in the United States
(DePaulo, 2006).
Despite these deeply ingrained beliefs about the importance of marriage for
mental health, recent evidence suggests a more nuanced perspective. The emerging
view among social scientists, based primarily on longitudinal research that follows
a group of individuals over time, is that the often-cited benefits of marriage and
the costs of marital dissolution are highly dependent on a range of individual
and contextual factors. These associations also appear to vary depending on which
group (divorced, widowed, never married) is compared to the married, how marital
status and living arrangements are defined (first married, remarried, cohabiting),
and the duration of time spent in particular statuses. This chapter examines the
evidence for an effect of marital status on mental health, with a particular focus
on the factors that identify who benefits from marriage, who suffers from marital
dissolution, and under what circumstances. Because several excellent reviews
of earlier literature exist (Avison, 1999; Ross, Mirowsky, & Goldsteen, 1990;
Umberson & Williams, 1999; Waite & Gallagher, 2000), we highlight research
published since 1999.
Measures of Mental Health
Studies of marital status differences in mental health commonly employ outcome
measures of psychological distress and depressive symptoms. Most have used the
Center for Epidemiologic Studies–Depression scale (CES-D), which measures
both emotional and physiological symptoms of depression. A substantial body of
research also has considered the importance of marital status for positive indicators
of subjective well-being, including happiness and life satisfaction, thereby recog-
nizing that mental health is more than the absence of distress (Diener, Lucas, &
Oishi, 2002; Keyes, 1998; Ryff, 1989). Although more limited in number, a few
studies have tested the association of marital status with specific diagnosable psy-
chiatric disorders such as schizophrenia, anxiety disorder, and major depression
(eg, Afifi, Cox, & Enns, 2006; Overbeek et al., 2006; Williams, Takeuchi, &
Adair, 1992a).
An additional outcome that has received increasing attention in recent years
is alcohol use or abuse. This reflects the recognition that men and women may
308 Kristi Williams, Adrianne Frech, and Daniel L. Carlson
express distress in different ways. A growing body of evidence suggests that
depressive symptoms and other internalizing or emotional reactions to stress are
more common among women, whereas externalizing or behavioral expressions of
distress such as alcohol abuse or even violence (Umberson, Williams, & Anderson,
2002) are more common among men (Horwitz & Davies, 1994). Thus, studies
that focus only on depressive symptoms may overstate the effects of marital status
on women’s mental health and underestimate its effects on men’s health.
Why Is Marriage Associated with Better Mental Health?
For more than a century (Durkheim 1897/1951), sociologists have argued that
social relationships, in general, and marriage, in particular, connect individuals to
society and to each other in ways that enhance mental health. Marriage is thought
to confer a sense of obligation, belonging, meaning, and purpose–aspects of social
integration that facilitate trust and the provision and receipt of emotional support
(Umberson & Williams, 1999; Waite, 2009). The economies of scale associated
with co-residence and the specialization that occurs in marriage provide economic
resources that are protective of mental health (Oppenheimer, 2000; Ross & Huber,
1985). The special legal status of marriage also confers a number of rights including
access to Social Security and public pensions, spousal health insurance benefits,
and the ability to make health care decisions for one another. The sense of security
provided by these spousal safety nets may have some positive consequences for
mental health.
The dominant explanation for the mental health benefits of marriage focuses
on these economic and psychosocial resources. At its most basic level, this mari-
tal resource model predicts that all unmarried individuals, including the never
married, divorced, and widowed, will have poorer mental health than the married
because they lack the health-enhancing resources provided by marriage. Similarly,
entering marriage should improve mental health, and exiting marriage should have
enduring negative consequences.
In contrast, the marital crisis model suggests that the strains of marital disso-
lution undermine mental health more than the resources of marriage protect it. If
this is true, there should be few differences in the mental health of the married
compared to the never married, who have never experienced marital dissolution.
Further, entering marriage should not substantially improve mental health. The
crisis model also generally predicts that exits from marriage should be associated
with transient declines in mental health from which individuals eventually recover,
although it remains possible that some strains associated with marital dissolution
and their effects on mental health persist.
Selection bias is an alternative explanation for marital status differences in
mental health. Selection bias attributes mental health differences between married
and unmarried persons to two sources. First, married individuals may have better
Marital Status and Mental Health 309
mental health to begin with: Rather than marriage causing improvements in well-
being, mentally healthier people may simply be more likely to enter into and remain
in marriage. This is essentially a problem of reverse causal order. Second, there may
be numerous other preexisting differences between married people and unmarried
people that cannot be completely measured by researchers. If those who choose
to marry (or divorce) differ from those who do not in characteristics that affect
mental health, associations of marital status with health may be biased or spurious
(Ribar, 2003). Because most sociological research relies on observational data,
selection bias is usually addressed through statistical correction or the inclusion of
control variables, with the goal of isolating the effect of marital status on mental
health. Researchers who study marital status differences in mental health cannot
use experimental research designs because this would mean that, based on random
assignment by the researcher, some people in the study would be required to get
married, whereas others would be required to stay single, divorce, or even become
widowed. Clearly people are not willing to change their marital statuses for the
sake of research, and asking them to do so would be highly unethical. The result,
though, is that married people differ from unmarried people in many different
ways, and this makes it difficult to determine if marriage causes mental health.
The majority of the historical evidence for a beneficial effect of marriage on men-
tal health is based on cross-sectional data collected at one point in time, which is
severely limited in its ability to differentiate among these competing explanations.
Certainly, cross-sectional studies that compare the married to an aggregate group
of all unmarried individuals cannot provide compelling evidence for one explana-
tion over another. Research that disaggregates the unmarried typically indicates
that marital status differences in mental health are greatest when the comparison
group is the divorced or widowed and smaller or nonsignificant when the compar-
ison group is the never married (Cairney & Krause, 2005; Williams et al., 1992a),
suggesting partial support for the crisis model. Most studies conducted in the past
10 years used longitudinal data that followed the same individuals over time to
more directly address these questions. The strengths of this research are that it
assesses how mental health changes in response to entrances into and exits from
marriage, considers whether changes in mental health associated with marital
transitions persist over time, and examines whether earlier mental health predicts
entry into or exit from marriage.
Marital Status and Mental Health: Recent Evidence
Marriage
Several recent longitudinal studies have demonstrated that entering marriage is
associated with increases in psychological well-being and declines in psycholog-
ical distress (Evans & Kelley, 2004; Frech & Williams, 2007; Horwitz, White, &
310 Kristi Williams, Adrianne Frech, and Daniel L. Carlson
Howell-White, 1996a; Kim & McHenry, 2002; Lamb, Lee, & DeMaris, 2003;
Marks & Lambert, 1998; R. Simon, 2002; Strohschein et al., 2005; K. Williams,
2003; Wu & Hart, 2002). These patterns appear to provide strong support for
the marital resource model. However, some scholars have raised the possibility
that improvements in well-being associated with entering marriage may be tem-
porary (ie, a honeymoon effect). This could explain why longitudinal studies
that span a relatively short time period find evidence of increased well-being
with entry into marriage, but cross-sectional studies (which include those in
longer term marriages) find smaller differences between the never married and
married.
Surprisingly little research directly tests this honeymoon effect for mental
health. Research on German adults indicates that life satisfaction increases in the
time surrounding marriage, but then declines again within about 5 years (Lucas,
et al., 2003; Zimmerman & Easterlin, 2006). After this point, scholars disagree
about whether life satisfaction completely returns to levels that existed prior to
marriage. Studies of changes in the quality of the marital relationship over the
life course have indirectly suggested that the mental health benefits of marriage
may dissipate with time. Marital quality appears to decline the longer people are
married (D. Johnson, Amoloza, & Booth, 1992; Umberson, Williams, Powers,
Chen, & Campbell, 2005; Van Laningham, Johnson, & Amato, 2001). This could
mean that the mental health benefits of marriage, which are dependent on marital
quality (Hawkins & Booth, 2005; K. Williams, 2003), also diminish over time.
Further, the evidence for even short-term improvements in mental health asso-
ciated with entry into marriage remains mixed. A few studies indicate no improve-
ment in mental health with entry into marriage (Hope, Power, & Rodgers, 1999;
Horwitz & White, 1991; Wu & Hart, 2002), and others suggest that “losing a
marital partner has a greater impact on psychological distress than gaining one”
(Strohschein et al., 2005, p. 2299), suggesting some support for the crisis model.
Several issues complicate comparisons across studies and are likely to be at least
somewhat responsible for discrepant findings. First, not all research distinguishes
entry into first marriage from remarriage. Studies focused specifically on remar-
riage suggest that it improves well-being relative to remaining divorced or wid-
owed (Barrett, 2000; Evans & Kelley, 2004; Kim & McHenry, 2002; Strohschein
et al., 2005; K. Williams, 2003), but conveys fewer mental health benefits than
first marriage (Barrett, 2000; Marks & Lambert, 1998). Combining the remarried
and the first married may, therefore, produce weaker associations of marriage with
mental health.
Second, variation exists across studies in the comparison group employed.
Arguably the most logical approach in determining whether marriage improves
mental health is to compare never-married individuals who transition into marriage
to those who remain continually never married. This is the design employed in
most studies, but others use a comparison group of the continually married or
Marital Status and Mental Health 311
a combined group of those in all marital statuses who experience no change.
Fewer differences can be expected when the newly married are compared to the
stably married, and greater differences may be seen when they are compared to
an aggregate group that includes the continually divorced and widowed.
Third, the choice of control variables and the timing of their measurement
are of great importance. Some studies ambitiously control for contemporaneous
changes or later values on a range of variables that are associated with mental
health, including income, social integration, and social support (eg, Wu & Hart,
2002). Such an approach can obscure a true effect of marital status if the control
variables are, in fact, mechanisms through which the mental health effects of
marital transitions are conferred (Lieberson, 1985). In other words, marriage may
improve mental health in part because it increases financial resources and gives
people greater access to sources of social support. If these factors are held constant
(essentially made equal for those who marry and those who do not), the association
of marriage with mental health is minimized. Variables such as income, financial
strain, emotional support, and others thought to mediate the marital status–mental
health association are generally best measured before the union transition of
interest or considered separately to test a possible mediating effect. As longitudinal
data create opportunities for increasingly sophisticated analyses, we must think
even more carefully about the implications of modeling strategies and study design.
The extent to which prior mental health selects individuals into marriage (ie,
reverse causality) remains unclear. Several studies suggest that neither depressive
symptoms nor alcohol abuse is associated with subsequent entry into marriage
(vs. remaining single; Kim & McHenry, 2002; Lamb, Lee, & DeMaris, 2003).
However, others indicate that, before marriage, those who eventually marry already
show lower psychological distress (Hope et al., 1999) and greater life satisfaction
(Lucas et al., 2003; Stutzer & Frey, 2006) than those who remain single. Most
scholars generally agree that selection explains some but not all of the association
between marital status and mental health. An advantage of longitudinal studies is
that, because they follow individuals over time, they can use statistical methods to
hold constant levels of mental health prior to marriage. Doing so allows researchers
to estimate the effect of entering marriage on mental health, assuming that those
who marry had the same prior levels of prior mental health as those who do not.
This reduces the role of selection in explaining observed associations of marital
status with mental health. However, other preexisting differences between the
married and the unmarried may still not be accounted for in this approach (see
Ribar, 2003).
Ultimately, determining whether marriage has a causal effect on mental health
requires a much more complete understanding of the processes through which any
effects are produced. Empirical evidence lags far behind theory in this respect.
The marital resource model suggests that marriage benefits mental health because
it provides economic and psychosocial resources, including social support and
312 Kristi Williams, Adrianne Frech, and Daniel L. Carlson
social integration, but few studies directly test this hypothesis. In many cases, the
existence of social integration or receipt of social support is simply equated with
occupying a particular marital status. Recent research seriously challenges this
approach. In contrast to long-held assumptions about the integrative functions of
marriage, Gerstel and Sarkisian (2006) reported that married people have less
contact and fewer support exchanges with parents, siblings, friends, and neighbors
than the unmarried (also see Sarkisian & Gerstel, 2008). This finding is consistent
with research on single mothers that finds they have more frequent contact with
family and friends and receive more support from them than do married mothers
(Marks & McLanahan, 1993). Future studies should continue to examine whether
entrances into and exits from marriage precipitate change in a range of personal
and social circumstances associated with mental health and identify the extent to
which these changes contribute to (or perhaps detract from) any mental health
benefits of marriage.
Divorce and Widowhood
Compared to the somewhat mixed findings for mental health benefits of enter-
ing marriage, research is much more conclusive regarding the costs of marital
dissolution. Numerous longitudinal studies have provided convincing evidence
that the transition to divorce is associated with increases in depressive symp-
toms and declines in well-being (Hope et al., 1999; Kalmijn & Monden, 2006;
Kim & McHenry, 2002; Marks & Lambert, 1998; Mastekaasa, 1995; Simon, 2002;
Strohschein et al., 2005; K. Williams, 2003; Wu & Hart, 2002). Widowhood, too,
is clearly accompanied by declines in mental health (Carr et al., 2000; Lee &
DeMaris, 2007; Lucas et al., 2003; Strohschein et al., 2005; Wade & Pevalin,
2004; Wilcox et al., 2003; K. Williams, 2003).
An unresolved issue is whether dissolution represents a temporary crisis from
which individuals recover or whether its negative consequences persist. One of the
earliest longitudinal studies to examine this question for divorce suggested that
individuals recover from the stress of marital dissolution after about 2 years (Booth
& Amato, 1991). More recent evidence, however, suggests that negative effects
persist despite some initial recovery (Hetherington & Kelly, 2003; Johnson & Wu,
2002; Lucas, 2005). In contrast, research on widowhood generally supports a crisis
model, indicating that mental health rebounds within a few years after the spouse’s
death (Harlow, Goldberg, & Comstock, 1991; Lee & DeMaris, 2007; Lund,
Caserta, & Dimond, 1993; McCrae & Costa, 1993; Mendes de Leon, Rapp, & Kast,
1994; Stroebe & Stroebe, 1993; Wilcox et al., 2003, but see Lucas et al., 2003).
Marital dissolution is best viewed as a process because its effects on mental
health unfold over time, even before the exit from marriage occurs. For example,
Lucas (2005) found that life satisfaction drops sharply in the 4 years preced-
ing divorce and reaches its lowest point about a year before the divorce occurs.
Marital Status and Mental Health 313
Therefore, support for the crisis model, which indicates that individuals return to
baseline mental health a few years after divorce, may be misleading if that baseline
is measured after the anticipatory increase in distress has already occurred (Lee &
DeMaris, 2007). In other words, divorced or widowed individuals may return to
their already elevated prior levels of distress, but still have worse mental health
than they did before their marital problems began. Although much progress has
been made in understanding the longitudinal course of marital dissolution, greater
attention to measurement timing combined with the increasing availability of data
spanning longer periods will help more fully address these questions.
Declines in well-being before marital dissolution also complicate tests of differ-
ential selection out of marriage. Several studies show that, before their marriage
ends, those who will subsequently divorce already have elevated distress levels
compared to those who do not divorce (Aseltine & Kessler, 1993; Hope et al.,
1999; DR Johnson & Wu, 2002; Lucas, 2005; Simon, 2002; Stutzer & Frey,
2006; Wade & Pevalin, 2004). This lower well-being of those who later divorce
compared to those who remain married may simply reflect the anticipation of
marital dissolution (Hope et al., 1999), rather than an effect of poor mental health
on the probability of divorce. Stronger support for selection is provided by a recent
study that demonstrates that those who later divorce are already less happy with
life before they even enter marriage than those who will not divorce (Stutzer &
Frey, 2006), but it is unclear whether this pattern exists for measures of psycho-
logical distress. Selection is generally thought to play less of a role in widowhood
as individuals have little control over this form of marital loss. Direct empirical
tests generally have found no evidence of lower well-being or elevated distress
several years prior to widowhood (Lucas et al., 2003; Wade & Pevalin, 2004, but
see Lee & DeMaris, 2007, and Wilcox et al., 2003).
Cohabitation
As cohabitation has outpaced marriage as the first union entry experienced by
young adults in the United States, its importance to mental health has received
growing attention. Compared to marriage, cohabiting unions are less stable
(Bumpass & Lu, 2000; Raley & Bumpass, 2003), offer fewer economic resources
(Brines & Joyner, 1999; Lerman, 2002), and have lower levels of relationship qual-
ity and commitment (SL Brown & Booth, 1996; Marcussen, 2005; Thomson &
Colella, 1992). Because of these differences, scholars have theorized that cohab-
itation should provide some but not all of the mental health benefits of marriage
(Ross, 1995). Cross-sectional research has generally supported this conclusion
with evidence that cohabitors are more depressed than their married counterparts
(Brown, 2000; Kim & McHenry, 2002; Marcussen, 2005; Wu, Penning, Pollard, &
Hart, 2003; but see Horwitz & White, 1998) but less depressed than unmarried
noncohabitors (Kurdek, 1991; Ross, 1995; but see Horwitz & White, 1998).
314 Kristi Williams, Adrianne Frech, and Daniel L. Carlson
Recent longitudinal evidence, however, calls into question the conclusion that
cohabiting unions offer any benefits for mental health. In several studies, entering
cohabitation is not associated with improved mental health over time (Kim &
McHenry, 2002; Lamb, Lee, & DeMaris, 2003; Musick & Bumpass, 2006;
Williams, Sassler, & Nicholson, 2008; Wu & Hart, 2002). However, the aver-
age associations that are examined may obscure important variation in the nature
and duration of cohabiting unions. Cohabitation is a precursor to marriage for
many adults: Half of cohabitors marry their partners (Bumpass & Lu, 2000). For
others, cohabiting unions function more as alternatives to singlehood than as trial
marriages (Manning & Smock, 2005) and appear to be motivated in part by eco-
nomic concerns and convenience (Sassler, 2004). A promising avenue for future
research is to consider whether long-term cohabiting unions that more closely
resemble or lead to marriage have different effects on well-being than those that
are short lived or are characterized by lower levels of commitment. Understand-
ing how different types of cohabiting unions affect mental health may also offer
insight into the more general mechanisms through which any protective benefits
of marriage are conferred.
Is Everyone Affected Equally by Marriage and
Marital Dissolution?
In sum, the best recent research suggests that, on average, entering marriage (but
not cohabitation) is associated with improvements in mental health and exiting
marriage is associated with declines in mental health, at least in the short run. One
of the most important–yet most frequently ignored–components of the preceding
statement is the phrase “on average.” Social science findings about the benefits of
marriage for mental health are based on average associations. Although it can be
tempting to conclude that averages apply to all or even most people, this is not
always the case. Averages are highly influenced by extreme cases (ie, outliers)
so that what appear to be substantial differences in mental health between the
married and the unmarried may in fact only apply to a small group of relatively
anomalous individuals. A growing body of research provides convincing evidence
that the mental health benefits of marriage and the costs of marital dissolution
vary greatly, depending on a range of individual, demographic, and relationship
characteristics. These are commonly referred to as moderators of the relationship
between marriage and mental health. We explore some of the most important
moderators in this section.
Gender
Of all of the moderators of the effect of marriage on mental health, gender has
received the greatest attention. In the early 1970s, sociologists argued that marriage
Marital Status and Mental Health 315
is less beneficial to women’s mental health than to men’s in part because women’s
marital roles are more demanding and require greater self-sacrifice (Bernard,
1972; Gove & Tudor, 1973). In fact, in describing what she called “his” and
“her” marriage, Jessie Bernard (1972, p. 37) wrote that “marriage introduced such
profound discontinuities into the lives of women as to constitute genuine emotional
health hazards.” Early studies generally supported the hypothesis that women
benefit less than men from marriage, but more recent research has challenged this
conclusion. Across a range of studies, the average mental health benefits associated
with entering marriage are similar for men and women (Evans & Kelley, 2004;
Kim & McKenry, 2002; Marks & Lambert, 1998; Simon, 2002; Strohschein et al.,
2005; K. Williams, 2003). Some studies indicated that men benefit more than
women from entering marriage in terms of reduced alcohol use (Marcussen, 2005;
Simon, 2002), but this benefit is balanced against women’s reductions in depressive
symptoms.
Interestingly, contemporary evidence for Jessie Bernard’s notion of “his” and
“her” marriage appears to be limited to remarriage following divorce or widow-
hood. The best evidence suggests that, compared to men, women receive fewer
mental health benefits from remarriage (Marks & Lambert, 1998; K. Williams,
2003). As K. Williams (2003) notes, these lower benefits may reflect the persis-
tence of gender roles that assign women primary responsibility for negotiating
complex interpersonal relationships between two families, a stressful process,
especially when children are involved (Whitsett & Land, 1992). Overall, though,
very little is known about men’s and women’s experiences in remarriage, and
some studies have found no gender differences in its mental health consequences
(Evans & Kelley, 2004; Strohschein et al., 2005).
Evidence for gender differences in the effects of marital dissolution on mental
health is mixed. With respect to divorce, several longitudinal studies suggest that
there is no significant gender variation in the effect of divorce on psychologi-
cal well-being (Kim & McKenry, 2002; Strohschein, et al., 2005; K. Williams,
2003) or alcohol use (Simon, 2002). Others, however, indicate that women expe-
rience greater increases in symptoms of depression and feelings of hostility, and
greater decreases in self-esteem, environmental mastery, and self-acceptance than
men after a divorce (Kalmijn & Monden, 2006; Marks & Lambert, 1998; Simon,
2002). Gender differences in the impact of widowhood on mental health also
depend in part on the outcome examined. Relative to the continually married,
men who become widowed report larger increases in depressive symptoms than
women, but women appear to experience greater losses to life satisfaction (Lee &
DeMaris, 2007; K. Williams, 2003). Even if the effects of marital dissolution are
similar for men and women, the explanations for these patterns likely differ at
least somewhat by gender. For example, marital dissolution appears to exact a
greater financial cost for women than for men, but men may face greater chal-
lenges in managing a household on their own or maintaining relationships with
316 Kristi Williams, Adrianne Frech, and Daniel L. Carlson
children (Peterson, 1996; Smock, 1994; Umberson & Williams, 1993; Umberson,
Wortman, & Kessler, 1992).
Marital Quality
Among the most important findings to emerge from the literature on marital status
and mental health in recent years is the observation that any psychological benefits
of marriage are highly dependent on the quality of the marital union. In fact, several
studies suggest that being in a strained, unhappy, or inequitable marriage is worse
for mental health than being single (Gove, Hughes, & Style, 1983; Hagedoorn,
Van Yperen, Coyne, & van Jaarsveld, 2006; Hawkins & Booth, 2005; K. Williams,
2003). Poor marital quality directly undermines mental health (Evans & Kelley,
2004; Frech & Williams, 2007; Hagedoorn et al., 2006; Hawkins & Booth, 2005;
K. Williams, 2003) and increases the risk of experiencing a range of mood, anxiety,
and substance disorders (Overbeek et al., 2006). Further, strained marriages do not
provide the same levels of emotional support and closeness as happy marriages,
and these are likely key resources through which any mental health benefits of
marriage are conferred. In sum, recent evidence strongly suggests that statements
about the mental health benefits of marriage should refer more specifically to the
mental health benefits of high - quality marital unions.
Because poor marital quality and divorce go hand in hand, it is important to
distinguish between their effects on mental health. Some evidence, for example,
has suggested that what appear to be negative effects of marital dissolution on
psychopathology are explained by the much stronger impact of the poor marital
quality that precedes divorce (Overbeek et al., 2006). Similarly, the extent to which
marital dissolution undermines mental health depends on predissolution marital
quality. The negative mental health consequences of divorce are notably weaker
for those exiting an inequitable or unsatisfying marriage (Hawkins & Booth, 2005;
Kalmijn & Monden, 2006; Wheaton, 1990; K. Williams, 2003). In fact, in some
cases a transition to divorce may be better for mental health–or, at least, appears to
be no worse–than remaining in a long-term unhappy marriage (Hawkins & Booth,
2005; K. Williams, 2003). Hawkins and Booth (2005, p. 445) sum up the evidence
as follows:“Unhappily married people may have greater odds of improving their
well-being by dissolving their low-quality unions as there is no evidence that they
are better off in any aspect of overall well-being than those who divorce.”
Age and Life Course
The life course perspective emphasizes that life course stage shapes the social
context and normative status of roles such as widowhood, divorce, or singlehood
(Bengtson & Allen, 1993; Elder, 1985) in ways likely to influence their effects
Marital Status and Mental Health 317
on mental health. Social stress theory has much in common with the life course
perspective in this regard as both emphasize the importance of the timing and
sequencing of transitions into and out of social roles (Pearlin & Skaff, 1996). Taken
together, these theoretical orientations suggest that several features of the life
course should be important to the effect of marital status on mental health, includ-
ing age, marital history, and the existence and age of children.
With respect to age, Marks and Lambert (1998) report that being unmarried and
exiting marriage more strongly undermine the well-being of younger compared
to midlife adults. Although being single is more normative at younger ages, this
pattern may reflect an accumulation of psychosocial resources by midlife that
facilitate coping with any strains of singlehood (Marks & Lambert, 1998). The
extent to which this is true, however, may also depend on prior marital history.
Barrett (2000) finds that second and, especially, third marriages offer fewer ben-
efits to mental health than first marriage. She also observes that the negative
consequences of divorce and widowhood are amplified among those who have
experienced these events previously. This latter finding is consistent with a life
course stress perspective that suggests that, over time, the accumulation of stress,
like the strains associated with marital dissolution, can make individuals more vul-
nerable to additional stressors (Ulbrich, Warheit, & Zimmerman, 1989). It is likely
that multiple dimensions of the life course interact in complex ways to impinge
on the association of marital status with well-being. For example, experiencing a
second divorce or entering a second marriage by age 30 is probably a qualitatively
different experience than doing so at age 50. Future research should consider how
multiple dimensions of the individual and marital life course work together to
influence well-being over time.
The sequencing and timing of key family roles like marriage and parenthood are
also important. For example, Williams and Dunne-Bryant (2006) find that the neg-
ative mental health consequences of divorce are greatest for those with preschool-
aged children and smaller or negligible for others. Greater parenting strains and
more frequent contact with the former spouse experienced by parents (especially
mothers) of young children appear partly responsible for this finding. These
results are noteworthy because only 20% of divorces in the United States involve
preschool-aged children (Fields & Casper, 2001; US Bureau of the Census, 1998).
For those divorcing later in the life course, divorce may have only a minor impact
on well-being. Similarly, Williams and colleagues (2008) find that entering and
exiting marriage has more negative effects for single mothers than for childless
women but that both groups of women experience similar mental health benefits
from marriages that endure. For single mothers, who tend to have a high risk of
divorce, entering marriage may pose some risks to mental health as entering and
exiting marriage are marginally worse for well-being than remaining single for
this group (Williams et al., 2008).
318 Kristi Williams, Adrianne Frech, and Daniel L. Carlson
Race/Ethnicity
To date, little research has considered whether the relationship between marital
status and mental health differs by race/ethnicity. This is an important and timely
question, especially given striking race/ethnic variations in patterns of family for-
mation. Although 90% of Americans will marry at least once, only two-thirds of
African Americans can expect to marry during their lifetime (Teachman, Tedrow, &
Crowder, 2000). The intersection of marriage and childbearing, too, varies sub-
stantially by race. In 2000, 27.1% of births to White women occurred outside of
marriage, compared to 42.7% of Hispanic births, and 68.5% of African American
births (Dye, 2005). That race/ethnicity is associated with such large variations in
patterns of marriage and family formation provides sufficient motivation to con-
sider how it might shape the consequences of marriage (and divorce) for mental
health and well-being. African Americans are also disproportionately affected by
government efforts to promote marriage among low-income single mothers, so
understanding the effects of marriage on the mental health of African American
women has substantial policy relevance.
The limited evidence available suggests that Whites benefit more from marriage
than other groups. Williams, Takeuchi, and Adair (1992) find that the cross-
sectional association of marital status with psychiatric disorder is much stronger
for Whites than for Blacks. In ongoing work, Kroeger and Williams (2007) observe
that a mental health advantage of marriage is largely limited to Whites and Black
men, with few marital status differences in mental health observed for Hispanics or
Black women. Other evidence indicates that Black mothers show better adjustment
to marital dissolution than their White counterparts (McKelvey & McKenry, 2000)
and that the stress associated with being single is significantly lower for Black
compared to White older adults (Pudrovska, Schieman, & Carr, 2006).
At this point we can only speculate about the why marital status may be more
important to the mental health of White compared to Black or Hispanic adults.
Nonmarital childbearing, long-term cohabitation, and permanent singlehood have
become increasingly normative experiences among African Americans (Edin &
Kefalas, 2005; Lichter, Qian, & Mellot, 2006; Sigle-Rushton & McLanahan,
2002). This likely shapes the meaning of being single for African Americans
in ways that are beneficial or, at least, less detrimental to mental health. Further,
some evidence has suggested that African Americans who marry experience lower
levels of marital quality than Whites (Broman, 1993; Goodwin, 2003), which may
attenuate any benefits of marriage or costs of dissolution. Race/ethnicity may also
influence the resources and strategies that individuals employ to manage stressful
marital experiences (Williams & Harris-Reid, 1999). As McKelvey and McKenry
(2000) illustrate, the better psychosocial adjustment to marital dissolution of Black
mothers as compared to White mothers results in part because of Black women’s
higher levels of personal mastery and utilization of formal instrumental support
Marital Status and Mental Health 319
after divorce. Family scholars have also made much of the role that extended
kin plays in African American and Hispanic families (Cherlin, 1998; Sarkisian &
Gerstel, 2004), and these sources of social support may serve some of the same
functions as marriage in protecting mental health.
Values, Beliefs, and Prior Mental Health
People enter marriage with a set of beliefs about the nature of the marital union
and their own level of commitment to its permanence. They also bring preexisting
mental health problems with them into marriage. Current research shows that these
factors strongly moderate the association of marriage and, in some cases, marital
dissolution, with mental health. For example, Simon and Marcussen (1999) find
that the negative effects of marital dissolution and the positive effects of entering
marriage are substantially greater for those with strong beliefs in the permanency,
desirability, and importance of marriage. Similarly, Booth and Amato (1991)
show that exiting marriage is most stressful for those who have high levels of
commitment to marriage.
Prior mental health history also plays a role in the benefits of entering a union.
Those who are depressed or least satisfied with life prior to marriage benefit most
from a transition into marriage (Frech & Williams, 2007; Lucas, 2005). In fact,
Frech and Williams (2007) find that entering marriage is associated with only
modest mental health benefits for those do not have symptoms of depression prior
to marriage–a group that makes up about 70% of their sample. The apparent
mental health advantage of entering marriage is approximately four times greater
for those who are highly depressed prior to marriage despite their lower levels of
marital quality (Frech & Williams, 2007). Although explanations for these patterns
await further investigation, highly depressed individuals likely have more to gain
than others from the companionship, emotional support, and stability of marriage.
It remains unclear, though, whether the spouses of previously depressed persons
benefit similarly to their partners or whether the average benefits experienced by
the depressed persist over time.
Conclusion
So, does marital status affect mental health? Although causality is difficult to deter-
mine, social science research on marriage and mental health suggests three general
conclusions. First, marriage does appear to offer some benefits to mental health,
at least in the short run. On average, those who enter marriage experience greater
improvements in well-being than those who remain single, providing support for
the marital resource model. Although prior mental health may play some role in
selecting individuals into marriage, it likely does not completely explain improve-
ments in psychological well-being associated with entering marriage. However,
320 Kristi Williams, Adrianne Frech, and Daniel L. Carlson
the duration of any mental health benefits of marriage and the specific mechanisms
through which such benefits are conferred remain somewhat unclear. It is also pos-
sible that the specific aspects of marriage that are most important to well-being
change over the life course.
Second, on average, exiting marriage appears to undermine mental health, but
at least some of these negative effects dissipate with time. It is likely that both the
marital resource model and marital crisis model are applicable to understanding
the impact of marital dissolution on psychological well-being. Clearly divorce
and widowhood involve numerous stressors, but the strong initial impact of many
of these stressors may be relatively short lived. Even after the dust settles and
individuals adjust to these stressful life events, the absence of several key resources
provided by marriage, such as emotional and financial support, may have some
enduring, albeit weaker, negative consequences for mental health.
Third, the average effects of marriage and marital dissolution on mental health
are far from universal. Although it can be tempting to conclude that the benefits of
marriage apply equally to all individuals, the best evidence suggests a great deal
of variation. Clearly, those with poor marital quality receive fewer if any mental
health benefits of marriage. Other factors that shape the meaning and experience
of marriage, including race/ethnicity, the presence and age of children, prior
mental health, and life course experiences, are also highly relevant. The result is
that social science research findings about the average mental health benefits of
marriage should never be used to guide decisions about whether an individual or
specific group of people should be encouraged to marry or remain married. And
headlines such as “New Treatment for Depression–Marriage”(MSNBC, 2007)
should obviously be viewed with a healthy dose of skepticism.
Finally, it is important to remember that marriage and marital dissolution both
involve a combination of rewards and strains. Most theorizing and research on
marriage focus on its rewards, whereas work on marital dissolution highlights its
strains. However, the extent to which either affects mental health likely depends
on the way in which the benefits stack up against the costs. We need to know much
more about the aspects of marriage that detract from its mental health benefits
and the positive changes that may follow from marital dissolution. The ability
to uncover these factors depends in part on the indicators of mental health that
we choose to examine. For example, research that examines other outcomes in
addition to depressive symptoms finds that the unmarried report greater levels
of personal growth and more personal autonomy than the married (Marks &
Lambert, 1998). As Marks and Lambert (1998, p. 676) conclude,“Marriage is
not a universal beneficial determinant of well-being. It appears wise, therefore, to
continue evaluating the effects of marriage on well-being with a multidimensional
lens whenever possible, so that we can obtain a more precise understanding of
how and when marriage is important for mental health.”